Site-Specific Modification of a Single-Chain Antibody Using a Novel

and. Department of Medicine University of Arizona, Tucson, Arizona. 85721 . Received September 29, 1998; Revised Manuscript Received February 25, 1999...
0 downloads 0 Views 213KB Size
Download PDF
424

Bioconjugate Chem. 1999, 10, 424−430

Site-Specific Modification of a Single-Chain Antibody Using a Novel Glyoxylyl-Based Labeling Reagent Zhan G. Zhao,†,‡ Jin S. Im,†,§ Kit S. Lam,*,†,§,| and Douglas F. Lake*,†,§ Arizona Cancer Center, 1515 North Campbell, Tucson, Arizona 85724, and Department of Microbiology and Immunology and Department of Medicine University of Arizona, Tucson, Arizona 85721

.

Received September 29, 1998; Revised Manuscript Received February 25, 1999

A novel, highly specific protein modification approach is described. By using conventional molecular cloning techniques, a protein can be constructed and expressed such that the N-terminal residue is replaced by cysteine. Its 1,2-aminothiol structure reacts very specifically with a glyoxylyl group at pH 7 or below, forming a relatively stable thiazolidine bridge. Therefore, a glyoxylyl-based labeling agent (e.g., radioactive tags, fluorescent probes, biotin) can be used to specifically modify a protein at its N-terminus. To highlight this novel approach, a recombinant anti-insulin single chain antibody (scFv) was specifically biotinylated at its N-terminus even in the presence of other proteins in the total cell lysate. The glyoxylyl-biotinylated scFv retained binding activity similar to unmodified scFv.

INTRODUCTION

One of the major efforts in bioconjugate chemistry research is to develop methods for the derivatization of macromolecules in a site-specific manner (1). Commonly, peptides and proteins are modified either on the side chains of lysine residues by using amine-reactive reagents or on the side chains of cysteine residues by using thiol-reactive reagents. Less popular methods include derivatization at the tyrosine residue side chains (2) and the enzyme-catalyzed reverse proteolytic reactions at the C-terminal carboxylate group (3). Among all these methods, only the enzyme-catalyzed reverse proteolytic reaction is truly site specific. Amine-reactive reagents, which include reagents based on active esters (e.g., N-hydroxysuccinimide ester), isothiocyanides, aldehydes, and sulfonyl halides, etc., react predominately with -amino functional groups and to a lesser extent R-amino functional groups. Despite the mild conditions under which these reagents are used, heterogeneity of the modified products arises because of the high abundance of lysine residues in proteins and many peptides. Thiol-reactive reagents, such as the maleimide-based reagents and the haloacetic-based reagents, are somewhat more selective because of the lower abundance of cysteine residues in peptides and proteins. In cases where all thiol groups are involved in disulfide bond formation, one can introduce a single surface-accessible thiol into the target protein by site-directed mutagenesis (4). Although thiol-reactive reagents are more specific than the amine-reactive reagents, reactions with other nonthiol nucleophiles are possible, especially when the modification is forced by use of excess maleimide-based reagents resulting in modification of the side chains of histidine, lysine, R-amino groups of peptides (5). An interesting approach to label only the N-termini was developed by Geoghegan et al. (6) and Gaertuer et * To whom correspondence should be addressed. † Arizona Cancer Center. ‡ Current address: SIDDCO, Inc., 9040 S. Rita Road, Suite 2338, Tucson, AZ 85747. § Department of Microbiology and Immunology. | Department of Medicine.

al. (7). This approach takes advantage of the special reactivity of IO4- toward vicinal -OH and -NH2 groups which exist in N-terminal serine or threonine. The 2-amino alcohol group can be easily oxidized by periodate at pH 7, generating a glyoxylyl group which specifically reacts with either a hydrazide group (6) or an aminooxy group (8). This approach has been used for specific labeling of interleukin-8 at the N-terminus with aminooxy-functionalized fluorescent probes (9) and poly(ethylene glycol) polymers (8). Later, Zhang and Tam (10) extended this approach by using a 1,2-aminothiol-based biotinylation agent which specifically react with the N-terminal glyoxylyl group forming a stable thiazolidine ring. To avoid potential oxidation of peptides and proteins by periodate, we developed an alternative approach which utilizes a glyoxylyl-based labeling agent to specifically label a synthetic peptide by reacting with either an N-terminal cysteine or a side-chain-attached cysteine resulting in a thiazolidine ring bridge (11). We also proposed that this approach could also be used to label a protein at its N-terminus where the original residue can easily be changed to cysteine using site-directed mutagenesis (11). In this paper, we report the cloning and expression of a recombinant protein with cysteine added to its N-terminus and the procedure of using a glyoxylylbased modification reagent to label this protein at the N-terminus either before or after its purification. To highlight this novel approach, we chose to modify a recombinant anti-insulin antibody, designated HB125 scFv1 (12), and a glyoxylyl-functionalized biotin as the modification agent. The conjugate products were then analyzed by Western blot in conjunction with streptavidin-alkaline phosphatase. A flexible and hydrophilic linker molecule was also synthesized and used in the synthesis of labeling agents in order to limit any unwanted structural perturbations of the protein being modified. MATERIALS AND METHODS

(+)-Biotin, 4,7,10-trioxa-1,13-tridecanedeamine, succinic anhydride, tetrakis(triphenylphosphine) palla-

10.1021/bc980120k CCC: $18.00 © 1999 American Chemical Society Published on Web 04/24/1999

Site-Specific Modification of a Single-Chain Antibody Scheme 1. Synthesis Fmoc-Ttds-OH

of

Linker

Bioconjugate Chem., Vol. 10, No. 3, 1999 425 Molecule,

dium(0), and sodium periodate were purchased from Aldrich (Milwaukee, WI); Fmoc-Lys(Aloc)-OH was from SNPE (Princeton, NJ); Boc-Ser(t-but)-OH and Fmoc-Osu were purchased from Advanced ChemTech (Louisville, KY). All other reagents and solvents were purchased from commercial sources and used without purification. ISMS and FABMS were acquired at the Mass Spectoscopy Facility, College of Pharmacy, University of Arizona. HRFABMS was acquired at the Mass Spectroscopy Facility, Department of Chemistry, University of Arizona. HPLC was carried out on an ISCO two-solvent system (Lincoln, NE). For analytical work, a column 250 × 4 mm i.d. (vydac) was used at a flow rate of 10 mL/min. Solvent A was prepared by adding 2 mL of TFA to 4 L of water. Solvent B was prepared by adding 1.7 mL of TFA to 500 mL of water and making up to 4 L with acetonitrile. Synthesis of Fmoc-Protected 4,7,10-Trioxa-1,13tridecanediamine Succinimic Acid (Fmoc-Ttds), (Scheme 1). 4,7,10-Trioxa-1,13-tridecanediamine (2.22 g, 10 mmol) was dissolved in 50 mL of CH3CN and placed in a 250 mL flask with magnetic stirring. Succinic anhydride (1 g, 10 mmol) in 25 mL of CH3CN was added dropwise over an hour. The reaction was allowed to proceed for an additional 3 h at room temperature. After the waxy product settled, organic solvent was decanted and discarded. The product was rediluted in 100 mL of 50% aqueous CH3CN and chilled in an ice bath for 30 min before Fmoc-OSu (4.4 g, 13 mmol) in 25 mL of CH3CN was added. Enough diisopropylethylamine was then added to maintain pH 8-9 throughout the reaction. After stirring for 10 h at room temperature, the solvents were removed in vacuo. The resultant product was dissolved in 100 mL of 5% NaHCO3 and washed with EtOAc. The aqueous phase was then acidified with 1 N HCl to pH 2 and extracted 3 times with 50 mL of EtOAc. The combined organic phase was washed with distilled H2O and dried over anhydrous Na2SO4. The evaporation of 1 Abbreviations: HBTU, O-benzotriazol-N,N,N′N′-tetramethyluronium-hexafluorophosphate; HOBt, N-hydroxybenzotriazol; DIEA, diisopropylethylamine; TFA, trifluoroacetic acid; ISMS, ion spray mass spectroscopy; FABMS, fast-atom bombardment mass spectroscopy; HRFABMS, high-resolutioin fast-atom bombardment spectroscopy; scFv, single-chain Fv; NCscFv, Nterminal cysteine-containing scFv; WTscFv, wild-type scFv; IPTG, isopropylthio-β-galactoside; NHS-LC-biotin, sulfosuccinimidyl-6-(biotinamido) hexanoate; BCIP, 5-bromo-4-chloro-3indolyl phosphate; NBT, nitroblue tetrazolium; IMAC, immobilized metal affinity chromatography; Ttds, 4,7,10-trioxa-1,13tridecanediamine succinimic acid.

Scheme 2. Sythesis of Glyoxylyl-Functionalized Biotin

a (i) AA, DIC, HOBt (AA ) Fmoc-Lys(Aloc)-OH, Fmoc-TtdsOH, Boc-Ser(t-But)-OH; (ii) 20% piperidine in DMF; (iii) tetrakis(triphenylphosphine) palladium(0); (iv) biotin, HTBU, HOBt, DIEA; v, 10% TFA in DCM; (vi) NaIO4 (0.02 M NaAc, 0.02 M EDTA, pH 4.2).

organic solvent gave 4 g of product (75% yield). Molecular weight calculated for this compound: 543.5. Found: 543.2 by ion spray mass spectroscopy (ISMS). Synthesis of Glyoxylyl Functionalized Biotin (Glyoxylyl-Biotin). Synthesis of the glyoxylyl functionalized biotin (glyoxylyl-biotin) is outlined in Scheme 2. Briefly, protected tripeptide, Boc-Ser-Ttds-Lys(aloc)resin, was synthesized with Fmoc solid-phase chemistry (13). Deprotection of lysine side chain, Lys(Aloc), was carried out by Loffexı´s procedure (14) with tetrakis (triphenyl phosphine) palladium(0) as the reducing agent. The on-bead biotinylation was performed by using HBTU and HOBt as the coupling agents. Enough (40-50%) diisopropylethylamine (DIEA) was added to dissolve biotin which is not soluble in pure DMF. The biotinylated tripeptide was then cleaved from the Rink resin with 95% TFA-5% H2O, and purified on preparative HPLC (ISCO, Lincoln, NE) using a linear gradient of 0 to 50% (by vol) B over 30 min. The product, Ser-Ttds-Lys(biotin)-NH2, has the expected molecular weight as determined by HRFABMS (M + H calculated for C33H61O10S 761.4231, found 761.4241). The biotinylated tripeptide was quantitatively oxidized following a published procedure (6). The resulting glyoxylyl functionalized biotin (glyoxylylbiotin) was purified by preparative HPLC. After loading the filtered reaction solution onto the column, the inorganic salts were eluted for 10 min with solvent A. The product was then eluted off with a linear gradient of 0 to 100% (by vol) B over 15 min. The molecular weight was determined by FabMS (M + H, 730.3, observed 729.7, and M + H + H2O, 748.3, observed 747.6). However, the signal was too small for a high-resolution MS analysis, probably because no ionizable group was present in the molecule. A similar result was also observed by Rose et al. (15). Construction of N-Terminal Cysteine Anti-Insulin scFv (NCscFv). Wild-type anti-insulin scFv(WTscFv) constructed from murine anti-insulin IgG1 (HB125) hybridoma, obtained from ATCC, was molecularly cloned into pET21d expression vector as previously reported (12). WTscFv was amplified from this vector using a 5′ primer which has a cysteine codon at the 5′ end (5′-

426 Bioconjugate Chem., Vol. 10, No. 3, 1999

Zhao et al.

Figure 1. Construction of N-terminal cysteine-containing scFv Vector preparation. Pet21d vector was digested with NcoI, followed by Klenow treatment to make blunt ends. After digestion with XhoI, the vector was purified. Insert preparation. WTscFv in pet21d was amplified with a 5′ primer hybridizing to the WTscFv with an N-terminal cysteine codon “TGC” and 3′ T7 terminator primer hybridizing to the vector. The resultant NCscFv was blunt ended with Klenow enzyme and purified. The NCscFv insert was digested with XhoI and purified. Vector and insert fragments were ligated together. The new NCscFv was then sequenced to ensure correct reading frame. Dotted lines represent nucleotide sequences.

TGCCAGATCCAGTTG) and a 3′ end T7 terminator primer (5′-GCTAGTTATTGCTCAGCGG) as shown in Figure 1. Briefly, the PCR product was blunt-ended with

Klenow enzyme for 30 min at 37 °C, electrophoresed on a 1% agarose gel, and purified using a Qiagen Gel Extraction Kit II. The N-terminal cysteine scFv (NCscFv)

Site-Specific Modification of a Single-Chain Antibody

insert was then digested with XhoI and purified. To prepare the vector for cloning, pET21d was digested with NcoI followed by Klenow treatment which resulted in blunt ends. After Klenow enzyme inactivation, pET21d was digested with XhoI. Finally, pET21d and the NCscFv insert were ligated overnight by T4 ligase. The ligation mixture was used to transform Novablue cells on an LB plate containing 50 µg/mL ampicillin. All molecular techniques were carried out according to Ausubel et al. (16). Fifty colonies were picked and grown in LB overnight after which minipreps were performed using a standard alkaline lysis procedure. Inserts were identified by PCR screening with 5′ T7 promotor and 3′ T7 terminator primers. Dideoxynucleotide DNA sequencing (17) was performed on selected clones to ensure correct position of the N-terminal cysteine codon at the 5′ end of the scFv. Expression and Preparation of Cell Lysates of NCscFv and WTscFv in Esherichia coli. BL21 (DE3) bacterial cells were transformed with pET21d containing either NCscFv or WTscFv genes. One colony of each transformed bacterial cells was grown separately in 50 mL of terrific broth (TB) at 37 °C, 200 rpm to OD ) 0.8. IPTG was then added to a final concentration of 1 mM, and the cells were grown for two more hours. Cells were harvested by centrifugation for 5 min in JA 17 rotor at 8000 rpm and resuspended in lysis buffer (50 mM TrisCl, pH 8, 1 mM EDTA, and 100 mM NaCl) containing 5 mg of lysozyme. After 1 h incubation at room temperature, the cells were sonicated three times for 15 s, then subjected to centrifugation at 8000 rpm for 5 min. The pellets of lysates were resuspended with 50 mM Tris-Cl, pH 8, and centrifugated. After resuspension in 8 M urea, 5 mM Tris-Cl, pH 8, lysates were sonicated again three times for 30 s each and incubated for 1 h on ice. Ureasolubilized cell lysates were collected after centrifugation. The supernatant was filtered through 0.45 µm membrane filter and saved for purification or direct cell lysate assay. Purification of NCscFv and WTscFv. The recombinant proteins were purified from the cell lysates on a nickel-sepharose column using previously reported methods (12). Fractions were collected and analyzed by SDSPAGE to determine which fraction contained desired proteins. Further purification of NCscFv for the purpose of protein sequencing was carried out by continuous elution SDS-PAGE gel electrophoresis using model 491 Prep Cell (Bio-Rad). Purified NCscFv was dialyzed against 25 mM triethylamine overnight, lyophilized, and resuspended in 100 µL of distilled water. NCscFv was precipitated at -70 °C overnight after the addition of 900 µL of acetone. The protein pellet was obtained by centrifugation at 13 000 rpm for 30 min and air-dried. The N-terminal cysteine residue of the purified NCscFv was derivatized with 4-vinylpyridine according to the protocol in the Applied Biosystems User Bulletin (no. 28, 1987), and the amino acid sequence was determined by an Applied Biosystems model 470A gas-phase sequencer equipped with a model 120 A path analyzer. Biotinylation of NCscFv and WTscFv in Cell Lysates. The prepared cell lysates were adjusted to pH 6 with 5× phosphate buffer (0.4 M NaOH, 0.2 M H3PO4, and 0.2 M boric acid). One microliter of glyoxylyl-biotin (0.4 mg/ml) or 1 µL of sulfo NHS-LC-biotin (0.4 mg/ml, Pierce, Rockford, IL) was added to 10 µL of cell lysates and incubated at room temperature for 30 min. The glyoxylyl-biotinylation reaction was stopped by adding 1 µL of L-cysteine (0.4 mg/mL) while the NHS-LC-biotin reaction was stopped by adding 1 µL of 150 mM TrisHCl buffer (pH 7.9).

Bioconjugate Chem., Vol. 10, No. 3, 1999 427

Biotinylation of Purified WTscFv, NCscFv, and Insulin. Purified protein was refolded using the same methods published before (12). WTscFv and NCscFv were finally dialyzed against carbonate buffer (15 mM Na2CO3 and 35 mM NaHCO3, pH 8). Sulfo NHS-LC biotin or glyoxylyl-biotin was added to WTscFv or NCscFv solution and incubated for 30 min at RT. The reaction was stopped by Tris HCl (1 M, pH 7.9) or free cysteine (1 mg/mL) for 2 h at RT. Further removal of free reagents from scFv solutions was done by dialyzing against carbonate buffer. Porcine insulin (Sigma, MO) was resuspended in PBS (pH 8.0) according to the manufacturer’s instruction and labeled with Sulfo NHS-LC biotin as described above. The insulin solution was dialyzed against PBS (pH 7.3) to remove free reagents. SDS-PAGE and Western Blotting. Biotinylated cell lysates and purified scFvs were mixed with 2× SDS sample buffer and loaded onto a 9% reducing SDS-PAGE gel. After electrophoresis, the gel was cut in half. Onehalf of the gel was stained with Coomassie Blue while the proteins on the other half of the gel were transferred to nitrocellulose paper in transfer buffer (25 mM Trisbase, 192 mM glycine, 20% methanol, and 1% SDS, pH 8.3). Transblotting was carried out at 30 V, 4 °C overnight using Trans Blot (Bio-Rad). Nonspecific binding to the nitrocellulose membranes was blocked by phosphatebuffered saline containing 0.1% porcine gelatin and 0.05% Tween-20 (blocking buffer) for 2 h at room temperature. Streptavidin-alkaline phosphatase (Pierce, rockford, IL) diluted in blocking buffer (1:5000 dilution) was incubated with the membranes for 1 h at room temperature. After washing with PBS-0.1% Tween, the membranes were developed by incubating with 5-bromo-4-chloro-3-indolyl phosphate (BCIP)/nitroblue tetrazolium (NBT), both obtained from Amresco (Solon, OH) in alkaline phosphatase buffer, pH 9.4 (1.5% diethanolamine, 0.5% NaN3, and 1 mM MgCl2) Insulin-Binding ELISA. A 96 well plate was coated with anti-insulin antibody HB125 (1 µg/mL) for 1 h at 37° and blocked with 3% BSA in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 4.3 mM NaH2PO4, and 1.4 mM KH2PO4, pH 7.3). Biotinylated insulin (10 nM) was preincubated with various concentrations of competitors such as WTscFv, NCscFv, or glyoxylylbiotinylated NCscFv for 1 h at RT. After HB125 coated plates were washed with washing buffer (0.05% Tween20 in PBS), preincubated biotinylated insulin and scFv mixtures were then added to HB125 IgG coated plate and incubated for another hour at RT. Unbound insulin, scFv and insulin bound scFv were washed out with washing buffer. Streptavidin-horseradish peroxidase (SA-HRP, Pierce,IL) was added and incubated for 1 h at RT to detect biotinylated insulin bound to HB125. 3,3′5,5′Tetramethylbenzidine base (TMB, Gibco BRL, MD) was added. Color development was stopped with 0.5 N H2SO4 and optical densities (OD) were measured at 450 nm. RESULTS AND DISCUSSION

Synthesis of Glyoxylyl Functionalized Biotin Reagent (Glyoxylyl-Biotin). One major concern for the synthesis of functionalized biotin (glyoxylyl-biotin) was whether the biotin group, which contains a thioether bond, remained unchanged when subjected to periodate oxidation. Fortunately, our results show that this is not a problem under the reaction conditions. Conventional cross-linking reagents contain a lengthy and flexible linker, but many of them are hydrophobic such as the hexanyl linker in NHS-LC-biotin. In this study we

428 Bioconjugate Chem., Vol. 10, No. 3, 1999

Zhao et al.

Table 1. Comparison between the Ligation Methods this paper site of derivatization protein oxidation with sodium periodate

N-terminal cysteine no

derivatizing reagent specific derivatization without protein purification

glyoxylyl compound yes

decided to use a lengthy, flexible, and hydrophilic molecule as the linker in order to minimize any unwanted structural perturbation of the target protein. For this purpose a poly(ethylene glycol) (PEG) linker would be ideal, but the commercially available PEGs usually consist of a mixture of PEG molecules with different molecular weights, thus causing problems in purification and characterization. Therefore, we chose a molecule which is similar to triethylene glycol and derivatized it into a special amino acid, so it can be used in standard Fmoc solid-phase peptide synthesis. As a result, the new biotin reagent shows excellent solubility in both aqueous and polar aprotic solvents. Our experiments also show that the glyoxylyl-based biotin reagent is very stable in both acidic and basic conditions (data not shown). This is clearly an advantage over other common NHS-based or maleimide-based biotinylating reagents. We are currently developing a solution-phase synthesis procedure, so that this glyoxylyl-based reagent can be made in bulk quantities. Site-Directed Mutagenesis and Expression of Anti-Insulin ScFv Containing N-Terminal Cysteine. As described in Materials and Methods, the codon for N-terminal cysteine was added by PCR-based sitedirected mutagenesis of WTscFv creating NCscFv. The NCscFv gene was then cloned into pET21d expression vector. Figure 1 shows the initial DNA sequences of the mutated fragments which contain codons corresponding to formyl-Met-Cys. Then pET21d vector containing NCscFv was used to transform BL21 (DE3) E. coli host cells for expression. Since the immunoglobulin fragments produced from the pET21d transformed cells contained a carboxyl (His)6, they were purified on a nickel-agarose column using immobilized metal affinity chromatography (IMAC) (18). Distinct bands at approximately 31 kDa can be seen in lanes 2 and 4 of Figure 2, corresponding to the wild-type scFv (WTscFv) and the N-terminal cysteine containing scFv (NCscFv), respectively. Purified antiinsulin NCscFv was subjected to protein microsequencing as described in Materials and Methods. Both DNA and protein sequencing analysis showed that a cysteine residue was successfully added to the N-terminus 5′ of glutamine, which was the first residue of WTscFv (data not shown). In our bacterial protein expression system, a methionine residue is incorporated at the N-termimus as a part of the universal translation-initiation signal used by prokaryotes as well as eukaryotes. In prokaryotes, the methionyl moiety carried by the initiator tRNA is Nformylated prior to its incorporation. For a significant fraction of intracellular proteins, the amino-terminal methionine is also removed enzymatically by a methionine-specific aminopeptidase after the initiation of translation (19, 20). We took advantage of these N-formyl methionine cleavage mechanisms as demonstrated by our ability to selectively label an N-terminal cysteine. Interestingly, methionine aminopeptidase has been reported to be particularly effective when cysteine is at the penultimate position (21).

Georghegan et al., Gaertuer et al., and Tam et al. (6, 7, 10) N-terminal threonine glyoxylyl group yes (convert N-terminal threonine or serine to glyoxylyl group) hydrazide or aminooxy compound to be proven

Figure 2. SDS-PAGE and Western blot Analysis of Derivatized proteins. Coomassie Blue stain of SDS/PAGE gel (lane M, 1-4) and Western blot of biotinylated and untreated lysates (lanes 5-10). Bands in lanes 5-10 were detected with streptavidinalkaline phosphatase using BCIP/NBT as substrates. Lane M, molecular weight marker; lane 1, WTscFv cell lysate; lane 2, purified WTscFv; lane 3, NCscFv cell lysate; lane 4, purified NCscFv; lane 5, sulfo NHS-LC biotin labeled WTscFv cell lysate; lane 6, sulfo NHS-LC biotin labeled NCscFv cell lysate; lane 7, glyoxylyl-biotin labeled WTscFv cell lysate; lane 8, glyoxylylbiotin labeled NCscFv cell lysate; lane 9, untreated WTscFv cell lysate; lane 10, untreated NCscFv cell lysate.

Biotinylation of Anti-Insulin ScFv in Total Cell Lysates. Total cell lysates that contained either WTscFv or NCscFv were biotinylated with (i) the conventional sulfo NHS-LC-biotin or (ii) our newly developed glyoxylylbiotin. The streptavidin-alkaline phosphatase system coupled with a colorimetric detection system (BCIP/NBT) was used to detect biotinylated products in the western blot (Figure 2). Biotinylation by sulfo NHS-LC-biotin was essentially identical for both WTscFv lysates (lane 5) and NCscFv lysates (lane 6). This is not suprising since NHSbased reagents label predominantly lysine side chains which are abundant in most proteins. Thus, the western blot pattern of lanes 5 and 6 resembled that of the Coomassie Blue stain as all proteins were labeled (compare lanes 5 and 6 with lanes 1 and 3). When glyoxylylbiotin was used to biotinylate total lysates, only one predominant band with a molecular mass of approximately 31 kDa in the NCscFv lysate was biotinylated (lane 8). This corresponds to the purified scFv shown in lanes 2 and 4. In contrast, only minor background was detected when the total WTscFv lysate was biotinylated with glyoxylyl-biotin under identical conditions (lane 7). As a negative control, no biotinylation or non-specific binding to streptavidin was detected for the untreated cell lysates (lanes 9 and 10). The above result clearly demonstrates that the glyoxylyl-biotin reagent is highly specific, that it can specifically label the N-terminal cysteine even in the presence of a large number of other proteins, and that glyoxylyl-biotin also works in the presence of 8 M urea under reducing conditions where many other free sulfhydryl groups of other proteins are exposed. Site-Specific Modification Does Not Significantly Alter the Binding Affinity of Anti-Insulin ScFv.

Site-Specific Modification of a Single-Chain Antibody

Figure 3. Insulin binding ELISA. A 96 well plate was coated with purified murin anti-insulin antibody (HB125) and blocked. Biotinylated porcine insulin was preincubated with serial dilutions of WTscFv, NCscFv, or glyoxylyl-biotinylated NCscFv, and was then place in HB125 coated 98 well plate. ScFvs and scFv0porcine insulin complex were removed by washing and biotinylated insulin captured by HB125 on plate was detected with SA-HRP and TMB [(9) WTscFv, (O) NCscFv, (b) glyoxylylbiotinylated NCscFv].

Purified anti-insulin WTscFv and NCscFv were refolded as described (12). To investigate the changes in binding affinities of N-terminally mutated anti-insulin scFv to insulin, an insulin binding ELISA was performed with WTscFv, NCscFv, and glyoxylyl-biotinylated NCscFv. First, a 96 well plate was coated with purified parent anti-insulin IgG antibody (HB125). Sulfo NHS-LC biotinlabeled insulin was preincubated with serial dilutions of WTscFv, NCscFv, or glyoxylyl-biotinylated NCscFv prior to adding to HB125-coated plate. During this preincubation step, scFvs bound biotinylated insulin. Since insulin has only one antigenic determinant recognized by parent HB125 anti-insulin antibody, insulin prebound to HB125 scFvs will no longer bind to plate-coated HB125 parent antibody. Thus, only free insulin will be captured by HB125 antibody while scFv-insulin complexes will remain in solution and washed away. Captured insulin was detected with SA-HRP and TMB substrate. The stronger the scFv binds to the insulin, the lesser amount of insulin will be captured by plate bound HB125 antibody, which in turn will lead to a lower OD 450 measurement. Using this assay, the relative reactivities of the three scFvs were assessed (Figure 3). All scFvs, NCscFv, WTscFv, and NCscFv-glyoxyl-biotin, inhibited insulin from binding the parent HB125 IgG in a concentration-dependent manner. WTscFv and glyoxylyl-biotinylated NCscFv similarly inhibit insulin from binding to the parent antibody, whereas the reactivity between insulin and underivatized NCscFv is slightly weaker than the other two in this competition assay. The above study indicates that site-specific modification does not significantly alter the binding affinity of scFv. Site-Specific Modification. There are several approaches available for site-specific protein modification (3, 6-8, 22, 23). Among them are the ones developed by

Bioconjugate Chem., Vol. 10, No. 3, 1999 429

Geoghegan (6), Gaertuer (7), and Tam (10). Their approaches are novel and easy to perform compared to other protein-labeling strategies. The strategy we present in this paper is partially inspired by their efforts and, therefore, shares some similarity with their approaches. For example, our method as well as the methods by Geoghegan et al. (6), Gaertuer et al. (7), or Tam (10) use glyoxylyl-based chemistry and both methods result in derivatization at the N-terminus of a protein. However, there are significant differences between our method and theirs (Table 1). For example, in our method, an N-terminal cysteine is incorporated into the protein using conventional site-directed mutagenesis. Then, the N-terminal cysteine, having a 1,2-aminothiol group, reacts specifically with glyoxylyl reagent forming a thiazolidine ring (11). Therefore, the protein itself is not modified after expression in the E. coli system. In contrast, the other methods require incorporation of N-terminal threonine or serine into the protein (by sitedirected mutagenesis) followed by oxidation with periodate, a condition that potentially can damage the protein. The second difference between the methods is that we introduced a glyoxylyl group into the derivatizing reagent whereas the other methods uses hydrazide, aminooxy, or 1,2-aminothiol groups. Finally, in our approach, the modification reaction can be performed in total cell lysates before proteins fold into their native forms, making this approach attractive for protein purification as well as detection. The reaction between glyoxylyl groups and N-terminal cysteines is extremely selective so that only the target protein is modified at its Nterminal cysteine, even in the presence of many other proteins and biological molecules (Figure 2). It is also important to note that the cell lysates are produced in bacterial cells in reduced form so that in the presence of >6 M urea, the N-terminal cysteine is free from forming disulfide bonds with other cysteine residues. Therefore, the presence of a 1,2-aminothiol structure is assured for the glyoxylyl-based ligation. Under the same conditions, other nonterminal cysteine residues are present in reduced form but are not modified by the glyoxylyl group as observed in our experiments. We believe this approach can be used as a general method to site-specifically label the N-terminus of any recombinant protein. One obvious application is to produce well-defined radioimmunoconjugates for radioimmunotherapy of cancer. Other applications include the selective labeling of sensitive proteins such as enzymes that could be inactivated by oxidation or over modification of free thiol or amino groups. LITERATURE CITED (1) Meares, C. F. (1993) Introduction to bioconjugate chemistry. In Perspectives in Bioconjugate Chemistry (C. F. Meares, Ed.) pp 1-8, American Chemical Society, Washington, DC. (2) Brinkley, M. (1993) A brief survey of methods for preparing protein Conjugates with dyes, haptens, and cross-linking reagents. In Perspectives in Bioconjugate Chemistry (C. F. Meares, Ed.) pp 59-70, American Chemical Society, Washington, DC. (3) Rose, K., Vilaseca, L. A., Werlen, R., Meunier, A., Fisch, I., Jones, R. M., and Offord, R. E. (1991) Preparation of welldefined protein conjugates using enzyme-assisted reverse proteolysis. Bioconjugate Chem. 2, 154-159. (4) Wingfield, P., Graber, P., Shaw, A. R., Gronenborn, A. M., Clore, G. M., and MacDonald, H. R. (1989) Preparation, characterization, and application of interleukin-1 mutant protein with surface-accessible cysteine residues. Eur. J. Biochem. 179, 565-571.

430 Bioconjugate Chem., Vol. 10, No. 3, 1999 (5) Smyth, D. G., Blumenfeld, O. O., and Konigsberg, W. (1964) Reaction of N-ethylmaleimide with peptides and amino acids. Biochem. J. 91, 589-595. (6) Geoghegan, K. F., and Stroh, J. G. (1992) Site-directed conjugation of nonpeptide groups to peptides and proteins via periodate oxidation of a 2-amino alcohol. Application to modification at N-terminal serine. Bioconjugate Chem. 3, 138-146. (7) Gaertner, H. F., Rose, K., Cotton, R., Timms, D., Camble, R., and Offord, R. E. (1992) Construction of protein analogues by site-specific condensation of unprotected fragments. Bioconjugate Chem. 3, 262-268. (8) Gaertner, H. F., and Offord, R., E. (1996) Site-specific attachment of functionalized poly(ethylene glycol) to the amino terminus of proteins. Bioconjugate Chem. 7, 38-44. (9) Alouni, S., Gaertner, H. F., Mermod, J. J., Power, C. A., Bacon, K. B., Wells, T. N. C., and Proudfoot, A. E. I. (1995) A fluorescent interlukin-8 receptor probe produced by targeted labeling at the amino terminus. Eur. J. Biochem. 227, 328334. (10) Zhang, L., and Tam, J. P. (1996) Thiazolidine Formation as a General and Site-Specific Conjugation Method for Synthetic Peptides and Proteins. Anal. Biochem. 233, 8793. (11) Zhao, Z. G., and Lam, K. S. (1995) A novel approach for iodolabeling synthetic peptides. J. Chem. Soc., Chem. Commun. 1739-1740. (12) Lake, D. F., Lam, K. S., Peng, L., and Hersh, E. M. (1994) Molecular cloning, expression and mutagenesis of an antiinsulin single chain Sv (ScFv). Mol. Immunol. 31, 845-856. (13) Atherton, E., and Sheppard, R. C. (1989) Solid-Phase Peptide Synthesis. A practical Approach, IRL Press, Oxford. (14) Loffex, A., and Zhang, H. X. (1992) New side-chain protection of amino acids: Potential use in SPPS. In Innovative and Perspectives in Solid-Phase Synthesis (Epton, R., Ed.) pp 7782, Intercept Limited, Anclover. (15) Rose, K., Zeng, W., Regamey P.-O., Chernushevich, I. V., Standing, K. G., and Gaertner, H. F. (1996) Natural Peptide

Zhao et al. as Building Blocks for the Synthesis of Large Protein-like Molecules with Hydrazone and Oxime Linkages, Bioconjugate Chem. 7, 552-556. (16) Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Sturuhl, K (1989) Current protocols in Molecular Biology, Creene Publishing Associates and Wiley-Interscience, New York. (17) Sanger, F., Nicklen, S., and Coulson A. R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. U.S.A. 74, 5463-5467. (18) Smith, M. C., Furman, T. C., Ingolia, T. D., and Pidgeon, C. (1988) Chelating peptide-immobilized metal ion affinity chromatography. J. Biol. Chem. 263, 7211-7215. (19) Meinnel, T., and Blanquet, S. (1995) Enzymatic properties of Escherichia coli peptide deformylase. J. Bacteriol. 177, 1883-1887. (20) Ben-Bassat, A., Bauer, K., Chang, S.-Y., Myambo, K., Boosman, A., and Chang, S. (1987) Processing of the initial methionine from proteins: Properties of the Escherichia coli methionine aminopeptidase and its gene structure. J. Bacteriol. 169, 751-757. (21) Hirel, P.-H., Schmitter, J.-M., Dessen, P., Fayat, G., and Blanquet. S. (1989) Extent of N-terminal methionine excision from Escherichia coli proteins is governed by the side-chain length of the penultimate amino acid. Proc. Natl. Acad. Sci. U.S.A. 86, 8247-8251. (22) Dı´Santos, C. S., Nicholson, G. C., Mosely, J. M., Evans, T., Martin, T. J., and Kemp, B. E. (1988) Biologically active derivatizable salmon calcitonin analog: Design, Synthesis and Applications. Endocrinol. 123, 1483-1488. (23) Wetzel, R., Halualani, R., Stults, J. T., and Quan, C. A. (1990) A general method for highly selective cross-linking of unprotected polypeptides via pH-controlled modification of N-terminal alpha-amino groups. Bioconjugate Chem. 1, 114-122.

BC980120K